app.py

import gradio as gr
import torch
import os
import numpy as np
from TTS.api import TTS
from scipy.io.wavfile import write

# Check if the user has agreed to the terms of service
tos_gr = gr.Checkbox(
    label="Agree",
    value=True,
    info="I agree to the terms of the CPML: https://coqui.ai/cpml",
)
tos_agreed = tos_gr.value

# Set the COQUI_TOS_AGREED environment variable
os.environ["COQUI_TOS_AGREED"] = "1" if tos_agreed else "0"

def generate_voice(text, language, sample_voice, tos_agreed):
    # Check if the user has agreed to the terms of service
    if not tos_agreed:
        return "Please agree to the terms of service before using the service."

    # Initialize TTS
    device = "cuda" if torch.cuda.is_available() else "cpu"
    tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2").to(device)

    # Load the audio sample from the user's upload
    speaker_wav_path = sample_voice.name

    # Use the sample voice to convert text to speech
    wav = tts.tts(text=text, speaker_wav=speaker_wav_path, language=language)

    # Save the generated audio to a temporary file
    output_path = "generated_audio.wav"
    write(output_path, 22050, np.array(wav, dtype=np.int16))  # Use np.int16

    # Return the path to the generated audio file
    return output_path

# Define the Gradio interface
iface = gr.Interface(
    fn=generate_voice,
    inputs=[
        gr.Textbox(lines=3, label="Enter Text"),
        gr.Dropdown(["es", "en", "fr"], label="Select Language"),
        gr.File(label="Upload Sample Voice"),
        tos_gr,
    ],
    outputs=gr.Audio(label="Generated Audio"),
)

# Launch the Gradio interface
iface.launch(share=True)